Swiftech H2O-120 Compact and Corsair Nautilus 500: Is Water Better?
by Wesley Fink on September 17, 2007 6:00 AM EST- Posted in
- Cases/Cooling/PSUs
New CPU Cooling Test Configuration
Cooling tests were first run on our existing test bed, which is the EVGA NVIDIA 680i SLI motherboard. Temperatures were measured using the NVIDIA Monitor temperature measurement utility, which is part of the nTune program.
All Cooling tests were then repeated on the new cooling test bed, which is a Rosewill R604-P-SL case sold without a power supply. The Rosewill is a www.Newegg.com brand typical of a moderately priced mid-tower case our readers might own. It was chosen because it is a Newegg top seller and includes a variable front intake louver and a quiet 120mm exhaust fan at the rear of the case. The case is also screw-less with components held in place by plastic holders instead of metal to metal connections. This appears to reduce case vibration and noise.
The new power supply is the Corsair HX620W, which has earned our respect as an exceptionally quiet PSU with a variable speed exhaust fan and a down facing intake fan mounted just above the CPU space in the case. To eliminate the video card as a source of noise we have moved to a fanless video card, currently a GeForce 7600 GS. Plans are to move to fanless DirectX 10 capable card in this test bed in the near future.
The new motherboard is the ASUS P5K Deluxe. This P35 chipset motherboard has exhibited outstanding overclocking capabilities in our testing, and it can also mount the newest 1333FSB Intel Core processors and can handle our existing high-speed DDR2 memory. The P5K3 uses heatsinks and heatpipes to cool board components so all motherboard cooling is passive.
Other components in the cooling test bed are generally the same as those used in our previous motherboard and memory test bed. We will feature the new test bed and its noise metrics in an upcoming cooling system review. The goal was to remove fans that might produce noise wherever possible, improve internal airflow, reduce the power supply noise floor, and eliminate the video card as a noise source. While we did add a 120mm exhaust fan to improve case airflow, the chosen exhaust fan is very low noise and will be replaced by the lowest noise 120mm fan when the 120mm fan roundup is completed.
All cooling tests are run with the components mounted in a standard mid-tower case. The idle and stress temperature tests are run with the case closed and standing as it would in most home setups. We do not use auxiliary fans in the test cooling case, except for the Northbridge fan attached to the 680i for overclocking. Room temperature is measured before beginning the cooler tests and is maintained in the 20 to 22C (68 to 72F) range for all testing.
Both Swiftech and Corsair provided a small tube of Arctic Silver Ceramic premium thermal grease with their water cooling kits. However, for consistency of test results we tested with our standard premium silver-colored thermal compound. In our experience the thermal compound used makes little to no difference in cooling test results. This is particularly true now that processors ship with a large manufacturer-installed heatspreader. Our current test procedure uses this standard high-quality silver-colored thermal paste for all cooler reviews.
We first test the stock Intel air cooler at standard X6800 speed, measuring the CPU temperature at idle. The CPU is then stressed by running continuous loops of the Far Cry River demo. The same tests are repeated at the highest stable overclock we can achieve with the stock cooler. Stable in this case means the ability to handle our Far Cry looping for at least 30 minutes.
The same benchmarks are then run on the coolers under test at stock, highest stock cooler OC speed (3.73GHz), and the highest OC that can be achieved in the same setup with the cooler being tested. This allows measurement of the cooling efficiency of the test unit compared to stock and the improvement in overclocking capabilities, if any, from using the test cooler.
The cooling test results of both water cooling kits are compared to a representative sample of air cooling results with the NVIDIA Monitor program on the EVGA 680i. This is done for a consistent comparison of water cooler test results to earlier air cooler results. We then repeated testing on the new test bed measuring results with both Intel TAT (Thermal Analysis Tool) and CoreTemp, a well-known and widely used shareware program for core temperature measurement. Both results are reported in additional test results so they can be compared and contrasted to each other and to NVIDIA Monitor results.
Noise Levels
In addition to cooling efficiency and overclocking abilities, users shopping for CPU cooling solutions may also be interested in the noise levels of the cooling devices they are considering. Noise levels are measured with the case on its side using a C.E.M. DT-8850 Sound Level meter. This meter allows accurate sound level measurements from 35bdB to 130dB with a resolution of 0.1dB and an accuracy of 1.5dB. This is sufficient for our needs in these tests, as measurement starts at the level of a relatively quiet room. Our own test room, with all computers and fans turned off, has a room noise level of 36.4dB. Procedures for measuring cooling system noise are described on page seven which reports measured noise results comparing the stock Intel cooler and recently tested CPU coolers to the Swiftech H2O-120 Compact and Corsair Nautilus 500 water cooling kits.
Cooling tests were first run on our existing test bed, which is the EVGA NVIDIA 680i SLI motherboard. Temperatures were measured using the NVIDIA Monitor temperature measurement utility, which is part of the nTune program.
All Cooling tests were then repeated on the new cooling test bed, which is a Rosewill R604-P-SL case sold without a power supply. The Rosewill is a www.Newegg.com brand typical of a moderately priced mid-tower case our readers might own. It was chosen because it is a Newegg top seller and includes a variable front intake louver and a quiet 120mm exhaust fan at the rear of the case. The case is also screw-less with components held in place by plastic holders instead of metal to metal connections. This appears to reduce case vibration and noise.
The new power supply is the Corsair HX620W, which has earned our respect as an exceptionally quiet PSU with a variable speed exhaust fan and a down facing intake fan mounted just above the CPU space in the case. To eliminate the video card as a source of noise we have moved to a fanless video card, currently a GeForce 7600 GS. Plans are to move to fanless DirectX 10 capable card in this test bed in the near future.
The new motherboard is the ASUS P5K Deluxe. This P35 chipset motherboard has exhibited outstanding overclocking capabilities in our testing, and it can also mount the newest 1333FSB Intel Core processors and can handle our existing high-speed DDR2 memory. The P5K3 uses heatsinks and heatpipes to cool board components so all motherboard cooling is passive.
Other components in the cooling test bed are generally the same as those used in our previous motherboard and memory test bed. We will feature the new test bed and its noise metrics in an upcoming cooling system review. The goal was to remove fans that might produce noise wherever possible, improve internal airflow, reduce the power supply noise floor, and eliminate the video card as a noise source. While we did add a 120mm exhaust fan to improve case airflow, the chosen exhaust fan is very low noise and will be replaced by the lowest noise 120mm fan when the 120mm fan roundup is completed.
| Cooling Performance Test Configuration | |
| Processor | Intel Core 2 Duo X6800 (Dual-core, 2.93GHz, 4MB Unified Cache) |
| RAM | 2x1GB Corsair Dominator PC2-8888 (DDR2-1111) |
| Hard Drive(s) | Hitachi 250GB SATA2 enabled (16MB Buffer) |
| Video Card | NVIDIA 7600GS (fanless) - All Standard Tests |
| Platform Drivers | NVIDIA 9.53 |
| NVIDIA nTune Intel TAT CoreTemp |
5.05.22.00 (1/16/2007) Version 2.05.2006.0427 Version 0.95 |
| Video Drivers | NVIDIA 93.71 |
| CPU Cooling | Swiftech H2O-120 Compact Corsair Nautilus 500 Thermalright Ultima-90 ZEROtherm BTF90 Xigmatek AIO (AIO-S800P) Evercool Silver Knight Enzotech Ultra-X 3RSystem iCEAGE Thermaltake Big Typhoon VX Thermaltake MaxOrb Scythe Andy Samurai Master Cooler Master Gemini II Noctua NH-U12F Asus Silent Square Pro Scythe Ninja Plus Rev. B OCZ Vindicator Thermalright Ultra-120 Extreme Thermalright Ultra-120 Scythe Infinity Zalman CNS9700 Zalman CNS9500 Cooler Master Hyper 6+ Vigor Monsoon II Lite Thermalright MST-9775 Scythe Katana Tuniq Tower 120 Intel Stock HSF for X6800 |
| Power Supply | Corsair HX620W OCZ PowerStream 520W |
| Motherboards | Asus P5K Deluxe (Intel P35) EVGA nForce 680i SLI (NVIDIA 680i) |
| Operating System(s) | Windows XP Professional SP2 |
| BIOS | EVGA Award P26 (1/12/2007) Asus AMI 0501 (06/26/2007) |
All cooling tests are run with the components mounted in a standard mid-tower case. The idle and stress temperature tests are run with the case closed and standing as it would in most home setups. We do not use auxiliary fans in the test cooling case, except for the Northbridge fan attached to the 680i for overclocking. Room temperature is measured before beginning the cooler tests and is maintained in the 20 to 22C (68 to 72F) range for all testing.
Both Swiftech and Corsair provided a small tube of Arctic Silver Ceramic premium thermal grease with their water cooling kits. However, for consistency of test results we tested with our standard premium silver-colored thermal compound. In our experience the thermal compound used makes little to no difference in cooling test results. This is particularly true now that processors ship with a large manufacturer-installed heatspreader. Our current test procedure uses this standard high-quality silver-colored thermal paste for all cooler reviews.
We first test the stock Intel air cooler at standard X6800 speed, measuring the CPU temperature at idle. The CPU is then stressed by running continuous loops of the Far Cry River demo. The same tests are repeated at the highest stable overclock we can achieve with the stock cooler. Stable in this case means the ability to handle our Far Cry looping for at least 30 minutes.
The same benchmarks are then run on the coolers under test at stock, highest stock cooler OC speed (3.73GHz), and the highest OC that can be achieved in the same setup with the cooler being tested. This allows measurement of the cooling efficiency of the test unit compared to stock and the improvement in overclocking capabilities, if any, from using the test cooler.
The cooling test results of both water cooling kits are compared to a representative sample of air cooling results with the NVIDIA Monitor program on the EVGA 680i. This is done for a consistent comparison of water cooler test results to earlier air cooler results. We then repeated testing on the new test bed measuring results with both Intel TAT (Thermal Analysis Tool) and CoreTemp, a well-known and widely used shareware program for core temperature measurement. Both results are reported in additional test results so they can be compared and contrasted to each other and to NVIDIA Monitor results.
Noise Levels
In addition to cooling efficiency and overclocking abilities, users shopping for CPU cooling solutions may also be interested in the noise levels of the cooling devices they are considering. Noise levels are measured with the case on its side using a C.E.M. DT-8850 Sound Level meter. This meter allows accurate sound level measurements from 35bdB to 130dB with a resolution of 0.1dB and an accuracy of 1.5dB. This is sufficient for our needs in these tests, as measurement starts at the level of a relatively quiet room. Our own test room, with all computers and fans turned off, has a room noise level of 36.4dB. Procedures for measuring cooling system noise are described on page seven which reports measured noise results comparing the stock Intel cooler and recently tested CPU coolers to the Swiftech H2O-120 Compact and Corsair Nautilus 500 water cooling kits.
Facebook
Twitter
RSS
58 Comments
View All Comments
EODetroit - Tuesday, September 18, 2007 - link
The above is the best comment here. In the end, its all about heat radiation. If the air coolers have a better means for radiating heat than the water coolers, they'll win. Water usually did well because it radiated the heat outside the case and most importantly, they could have bigger radiators than the little HS&F inside the case.With the modern high end air coolers, the radiators are huge (even if they're inside the relatively hot case). With the water cooler setups you tested, the radiators are small (small for a water radiator, even if its radiating to the outside air). Therefore you don't really need to see the numbers to know what the result will be.
To take advantage of water cooling, you need a big time radiator on the outside of the case... preferably with large fans that can push a lot of air through them without making a lot of noise. By testing radiators the size of a 120mm fan, you negate water cooling's biggest advantage.
Lem - Wednesday, September 19, 2007 - link
I kinda agree but...I somewhat disagree with the reasoning because heatpipes transfer heat a lot better than the water tubes we see used with water cooling systems.
The only benefit of water cooling is basically the fact that you can transfer heat longer distances. This allows you to take advatage of big external heat sinks.
If there were long and flexible heatpipes out there, we would not even consider water.
GhandiInstinct - Monday, September 17, 2007 - link
We need more important articles like these, as well as DDR2 comparisons and GPU comparisons....Anyway, I have one minor complaint, I wish the line graphs would be done away with as well as maybe labeling the names with a (AIR) and (WATER) so we know which performs best :)
(It's very hard to tell with the line graph)
Jethrodood - Monday, September 17, 2007 - link
I cant believe such a long standing tech site would draw such naive conclusions in the realm of cpu cooling. /scratches head..JAG87 - Monday, September 17, 2007 - link
This has to be one of the worst articles ever. It was decent until the very last page, where conclusions are drawn based on two cheap beginner water cooling kits, and one test platform.First, don't say water cooling is bad when testing beginner All-in-one test kits. These are cheap, and they perform as much as they cost.
And second, try keeping an over clocked quad core below 60c at load with air cooling. Take a 3.6 to 4 Ghz quad, thats 200W of heat. Try cooling that with air, just try.
And you don't have to spend more than 300 for water cooling to beat air cooling. Example: Apogee GTX/Dtek Fuzion + MCR220/MCR320 + MCP655 + MCRES MICRO + 7/16ID tubing + a few zip ties = way less than 300 dollars.
Why don't you split your review into pricing categories, and then you can draw a conclusion for each price category, instead of calling shens on water cooling. Make a sub 200 dollar section, a 200-300 dollar section, and a 300+ section, and run tests for each. Then you can say, hey sub 200 kits are a waste of money, 200-300 kits are the sweet spot for CPU cooling, and 300+ kits are only for people who wish to water cool more than just the CPU.
Come on Wesley, we expect better from Anandtech.
oopyseohs - Monday, September 17, 2007 - link
And second, try keeping an over clocked quad core below 60c at load with air cooling. Take a 3.6 to 4 Ghz quad, thats 200W of heat. Try cooling that with air, just try.QFT
I am by no means an advocate of water cooling, but I do appreciate the ability of water cooling to dissipate greater quantities of heat faster than air cooling. Maybe not faster (which is why you're getting better scores on a low TDP X6800), but more. I have tested the H20-120 Compact on QX6850's with Gabe and it's superiority to even the best air coolers is quite evident - even though it has just a single 120mm radiator.
While I don't think this is "one of the worst AnandTech articles ever", as a previous poster so delicately put it, I think it does bring to light an important flaw in your testing methodology for coolers: that the Core 2 Duo just does not generate anywhere near the heat of the increasingly-common Core 2 Quad.
strikeback03 - Wednesday, September 19, 2007 - link
I'm currently typing on a system with a Q6600 on an MSI P35 Platinum board with the jumpers set to the 333 bus speed, giving a 3.0GHz quad core. Tuniq tower at around 1400RPM, running quad Prime95 Speedfan shows core temps in the mid 50*C range.rallycobra - Monday, September 17, 2007 - link
does anyone have actual temp data to support this? At some point does the aircooling show a non linear response to additional heat and the temp shoots much higher than water cooling? I'm not sure how this is possible.rallycobra - Monday, September 17, 2007 - link
I disagree. This is an excellent article. An $80 Thermalright 120 ultra has better thermals/noise/cost/install. And no chance of a catastrophy with fluid! If the data is reliable, and I believe it is, there is no sense in anyone spending $150 on either of these kits.I was thinking of getting the swiftech kit, since other reviews from less credible sources said it was better than any air solution, and then upgrading the radiator to a thermochill 120.3 and keeping the apogee drive. That is a $150 radiator with a $100 cpu block/pump.
It appears to me that you need to budget about $300 for fluid to be better than air. If you are going to go through the effort, go big or don't bother. I'm talking a huge radiator. If swiftech releases a kit with a 120.2 radiator for the same price, that may make sense.
Calin - Monday, September 17, 2007 - link
Because this was a review of a pre-made, as-sold water cooling setup. You mentioned he compared installation time with the air coolers?The $50 and 15 minutes of installation for a top-end air cooler, compared to the hassle of buying all the afore-mentioned components and fitting them, for a $300 price point (maybe?)
I'd say he got the right conclusion, and the article was sound. It was not intended to the water cooling experts, but to people that would (maybe) choose a somewhat simple to install $150 solution instead of a simple to install $50 solution.
By the way, in all the list you put (Apogee GTX/Dtek Fuzion + MCR220/MCR320 + MCP655 + MCRES MICRO + 7/16ID tubing + a few zip ties), I only know for sure what zip ties are.